Heuristic methods of artificial intelligence - a new approach to line identification in vibrational-rotational spectra
A.P. Shcherbakov, O.V. Naumenko, A.D. Bykov
V.E. Zuev Institute of Atmospheric Optics of Siberian Branch of the Russian Academy of Science, Tomsk, Russia
Keywords: vibrational-rotational spectra, potential function method, neural network, perceptron, effective Hamiltonian, dipole moment
Abstract
This article addresses the problem of line identification in high-resolution vibrational-rotational spectra. We suggest an approach based on heuristic search applied in artificial intelligence systems for automatically identifying a set of molecular energy levels observed in a spectrum. Such systems solve the problem of finding a sequence of operations for transforming input data (formulas, knowledge) that brings the data to a certain target or desired state. The proposed approach simultaneously analyzes several rotational sublevels in a spectrum, thus enabling more reliable identification of the corresponding spectral lines using more information. The results can be used in optical atmospheric measurements and molecular physics.
|
